7326
Biochemistry 1989, 28, 7326-7332
a-Thrombin-Catalyzed Activation of Human Platelet Factor XIII: Relationship between Proteolysis and Factor XIIIa Activity? Thomas J. Hornyak,' Paul D. Bishop,§ and Jules A. Shafer*,t Department of Biological Chemistry, The University of Michigan, Ann Arbor, Michigan 48109, and ZymoCenetics, Inc., 4225 Roosevelt Way N E , Seattle, Washington 98105 Received January 31, 1989; Revised Manuscript Received May 16, 1989
ABSTRACT: The kinetics of activation of platelet factor XIII, an a-subunit dimer, were characterized by determining rate constants for activation peptide (AP) release, generation of activity, and exposure of the active-site thiol group. The specificity constant (kat/&,) for a-thrombin-catalyzed A P release, 1.2 X lo5 M-' s-l, was found to be similar to that for A P release from the tetramer plasma factor XIII (a2b2)[Janus, T. J., Lewis, S. D., Lorand, L., & Shafer, J. A. (1983) Biochemistry 22, 6269-62721, implying that the b subunits of plasma factor XI11 do not hinder a-thrombin-catalyzed cleavage of A P from the a subunit. Platelet factor XIIIa activity was generated a t a rate approximately twice the rate of A P release. This difference in rates was shown to be consistent with a reaction pathway for activation of platelet factor XI11 wherein full factor XIIIa activity is generated when one A P is removed from the dimeric zymogen so that removal of the second A P has no detectable effect on catalytic activity. In accord with this conclusion, the rate constant for exposure of the active-site thiol group, as measured by the incorporation of [ 1-14C]iodoacetamide, was about twice that observed for the removal of AP. The observed stoichiometry of labeling with iodoacetamide of only one active-site thiol group per dimer for active factor XIIIa, together with the values of the rate constants for A P release, generation of activity, and exposure of the active-site thiol group, suggests that either (i) release of one molecule of AP per platelet factor XI11 dimer triggers subunit interactions that cause both active-site thiol groups to become exposed with one thiol group becoming unreactive upon alkylation of the other or (ii) release of one molecule of A P per platelet factor XI11 dimer triggers subunit interactions that expose one active-site thiol group and prevent simultaneous exposure of both active-site thiol groups even after the second molecule of A P is released.
F a c t o r XIIIa, the last enzyme in the blood coagulation cascade, is a transamidase that catalyzes formation of yglutamyl-t-lysyl peptide cross-links between polypeptide chains in adjacent fibrin monomers and between fibrin and other plasma proteins [for reviews, see Lorand et al. (1980) and Shafer and Higgins (1988)l. Cross-links between fibrin molecules increase the mechanical strength of fibrin clots (Roberts et al., 1974; Gerth et al., 1974), whereas the crosslinks between fibrin and adhesive glycoproteins such as thrombospondin and fibronectin may serve to anchor fibrin clots to the site of injury (Shafer & Higgins, 1988). Additionally, the cross-linking of fibrin to a2-antiplasmin which occurs during blood clotting increases the resistance of fibrin clots to the fibrinolytic activity of plasmin and appears to be an important determinant of the lifetime of fibrin clots (Sakata & Aoki, 1980, 1982; Jansen et al., 1987). Factor XIII, the zymogen form of factor XIIIa, is found in the body in two forms. In plasma, factor XI11 exists as a tetramer comprised of two a subunits and two b subunits. A homodimeric form containing only the a subunits of plasma factor XI11 is located within platelets and placental tissue (Bohn, 1972; Schwartz et al., 1973, 1974). The two forms of factor XI11 (a2b2and az)are both activated by a-thrombin
Scheme I
'This research was supported by National Institutes of Health Grant H L 32006 (to J.A.S.). T.J.H. has received support from National Institute of General Medical Sciences Grant T32 GM07863 of the Medical Scientist Training Program and is currently a recipient of the Young Scientist Training Program M.D./Ph.D. scholarship from the Life and Health Insurance Medical Research Fund. *The University of Michigan. 5ZymoGenetics, Inc.
Abbreviations: AP, activation peptide; BPB, bromophenol blue; DTT, dithiothreitol; EDTA, ethylenediaminetetraacetic acid; HEPES, N-(2-hydroxyethyl)piperazine-N'-2-ethanesulfonic acid; PEG, poly(ethylene glycol); PPACK, D-phenylalanyl-L-prolyl-L-arginine chloromethyl ketone; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; r-factor XIII, yeast recombinant human platelet factor XIII; TANEP, 0.1 M Tris-acetate, 0.15 M NaCI, 1 mM EDTA, and 0.1% PEG, pH 7.5; Tris, tris(hydroxymethy1)aminomethane.
0006-2960/89/0428-7326$01 .50/0
ca2+
plasma factor X I I I : a2bz AP
platelet factor XIII:
a2*
b2
a2 thrombin
AP
as shown in Scheme I. a-Thrombin acts upon both plasma and platelet factor XI11 to release a 37-aminoacyl N-terminal peptide, activation peptide (AP),' from the amino terminus of the a subunits (Takagi & Doolittle, 1974; Ichinose et a]., 1986). In the presence of calcium ion, the b subunits dissociate from a-thrombin-treated plasma factor XI11 (Lorand et al., 1974; Cooke & Holbrook, 1974; Chung et al., 1974), leaving a cleaved dimer (a2)identical with the product of thrombin action on platelet factor XIII. In the continued presence of calcium ion, the active-site thiol group is unmasked (Curtis et al., 1973, 1974; Cooke & Holbrook, 1974; Chung et al., 1974), resulting in the conversion of a2/ to a2*,the catalytically competent form of factor XIIIa. Activation of plasma factor XI11 is a highly regulated process, with fibrin serving as a cofactor both in the Ca2+-
'
0 1989 American Chemical Society
Activation of Platelet Factor XI11 dependent conformational change necessary for exposure of the active-site thiol (Credo et al., 1978, 1981) and in the initial proteolytic event (Janus et al., 1983; Lewis et al., 1985). In two previous studies comparing activation of plasma and platelet factor XI11 (Chung et al., 1974; Greenberg et al., 1987), it was proposed that the b subunits inhibit access of a-thrombin to the scissile bond at Arg-a37. In one of these studies (Greenberg et al., 1987), it was also proposed that fibrin interacts with plasma factor XI11 to remove the inhibition of a-thrombin-catalyzed cleavage by the b subunits. The athrombin-catalyzed initial proteolytic event, cleavage of AP, was not directly measured in these studies. Instead, the studies relied upon the measurement of factor XIIIa activity to provide information about the initial proteolytic cleavage event. The possible complications associated with such a determination prompted us, in the present study, to characterize directly the kinetics for the a-thrombin-catalyzed event in the activation of platelet factor XIII. Comparison of these kinetics to those previously determined for plasma factor XI11 (Janus et al., 1983) indicates that the b chains of plasma factor XI11 do not alter the access of a-thrombin to the scissile bond at Arga37. Additional findings presented in this paper show that release of only one AP per dimer fully activates platelet factor XIII.
MATERIALS AND METHODS Platelet-Derived Factor XIII. Platelet factor XI11 was purified from 3-week-old human platelets by using a published procedure (Schwartz et al., 1974) with small variations. Instead of sonication, a Sorvall Omni-mixer (setting 5 , 4 X 10-s bursts) was used to lyse the platelets. Gel filtration chromatography was conducted using Sephacryl S-300 in 50 mM HEPES-NaOH, 0.15 M NaCl, and 1 mM EDTA at pH 7.5. Fractions of highest specific activity were concentrated by ultrafiltration under N 2 gas. When necessary, a second chromatography using DE-52 or a second 40% ammonium sulfate precipitation was added to purify further the platelet factor XIII. Platelet factor XI11 purified by this method was estimated to be >90% pure by SDS-PAGE and Coomassie staining. The purified platelet factor XI11 was stored at 4 OC in 50 mM HEPES-NaOH, 0.15 M NaCl, and 1 mM EDTA, pH 7.5, and retained 87% of its activity under these conditions after 9 weeks. Yeast Recombinant Human Platelet Factor XIII (r-Factor XIII). Factor XI11 was subcloned as a PstI fragment from a full-length cDNA isolated by Ichinose et al. (1986). The cDNA was expressed by joining to the yeast ADH2-4c promoter (Russell et al., 1983) and the TPIl terminator (Alter & Kawasaki, 1982). This expression cassette was subcloned into a yeast 2p-based expression vector containing the Schizosaccharomyces pombe triosephosphate isomerase (TPI) gene, and the construction was transformed into the tpi yeast strain ZM118, genotype ( a l a pep4::URA3/pep4::URAS; leu2-3,112/leu2-3,112 Atpi::URA3/Atpi::URA3barl-11 barl-1 [ c i r o ] ) .A 12-L fermentor was inoculated with yeast and fed glucose continuously until the absorbance reached 30-45 (at 600 nm). The glucose feed was then stopped, and the cells were fed continuously with 2% ethanol for the duration of the fermentation, usually 24 h. Unless otherwise stated, all buffered solutions contained 5 mM EDTA and 5 mM 2-mercaptoethanol, Whole cells were harvested, solvent-exchanged, and concentrated by using a spiral cartridge system (Amicon, Danvers, MA) to yield a final concentration of 50% (packed wet weight to volume) in deionized H 2 0 . The resulting cell suspension was then diluted with concentrated lysis buffer to give a final cell concentration of 40% in 50 mM Tris-HC1, pH 7.4, 150 mM NaCl, 15 mM
Biochemistry, Vol. 28, No. 18, 1989
7327
EDTA, 5 mM 2-mercaptoethanol, and 1 mM phenylmethanesulfonyl fluoride. Cell lysis was performed in a glass bead mill (DynaMill, Glen Mills, MA) using 100-pm acidwashed glass beads. Cell debris was removed from the lysate by centrifugation at 4000g for 45 min. The lysate was further clarified by addition of streptomycin sulfate to a final concentration of 2% (w/v) followed by centrifugation at 9000g for 60 min. The resulting supernatant was fractionated by addition of PEG-8000 to 12% (w/v) and the factor XI11 containing precipitate harvested by centrifugation. This crude factor XI11 precipitate was then redissolved in 50 mM TrisHC1, pH 7.8, and applied to a 6 X 27 cm (diethylaminoethy1)cellulose fast-flow Sepharose (Pharmacia LKB, Piscataway, NJ) column, washed with 2 column volumes of the same buffer, and then eluted with a linear gradient consisting of 1 L each of the following: buffer A-50 mM imidazole, pH 6.3; buffer B-50 mM imidazole, pH 6.3, and 150 mM NaCl. The factor XI11 containing peak fractions were then pooled, and the factor XI11 was precipitated by dialysis against 50 mM piperazine, pH 6.0. The resulting crystalline precipitate was then redissolved in 50 mM glycine, pH 7.6, and chromatographed on a 4.7 X 80 cm S-200 Sephacryl (Pharmacia LKB, Piscataway, NJ) column. The peak fractions containing factor XI11 were then pooled, filter-sterilized, and either used directly or stored as piperazine precipitate at 4 OC. The resulting product appeared to be >99% pure as judged by silver-stained SDS-PAGE, by gel filtration chromatography, and by the virtual absence of any detectable yeast protein in antiyeast Western blots. Yeast recombinant human factor XI11 (r-factor XIII) appears to be indistinguishable from placental or platelet-derived human factor XI11 by all of the criteria thus far studied. Like the placenta and platelet-derived factor XIII, native r-factor XIII (i) is a homodimer with an M , of 165 000 as determined by equilibrium sedimentation (David C. Teller, unpublished observations), (ii) is blocked to N-terminal Edman degradation, and (iii) yields upon treatment with a-thrombin a 4-kDa N-blocked activation polypeptide and a 79-kDa active protein (factor XIIIa). Additionally, the N-terminal sequence of the first 20 aminoacyl residues (as determined by Edman degradation) of a-thrombin-generated r-factor XIIIa is identical with that of platelet and placenta-derived factor XIII, and all three proteins catalyze the rapid formation of fibrin cross-links, yielding a similar pattern of y dimers and a multimers when analyzed by reducing SDS-PAGE. Platelet Factor XIIIa Assays. Concentrations of platelet factor XI11 were determined from the absorbance of stock solutions using a value of 13.8 for El& and a molecular weight of 150000 (Schwartz et al., 1973; Chung et al., 1974). Assays of platelet factor XIII, both for quantifying platelet factor XI11 in fractions during purification and for measuring the activity of purified enzyme, were performed by using a modification of a previously described procedure (Curtis & Lorand, 1976). A solution prepared by mixing 20 pL of 4 mM dansylcadaverine in 50 mM Tris-HC1, pH 7.5, 50 pL of 0.4 M CaC12 in 50 mM Tris-HC1, pH 7.5, 200 pL of the platelet factor XI11 containing sample, and 1200 pL of 50 mM Bicine-NaOH, pH 9.0, was equilibrated at 37 O C in a cuvette. The solution was mixed with 50 pL of 500 units/mL a-thrombin in 25 mM Tris-HC1/25% glycerol, pH 7.5, and incubated for 10 min to activate the factor XIII. A 50-pL aliquot of 0.2 M DTT in 50% glycerol was then added, and after 1 min, 200 pL of 2% A’,”-dimethylcasein in 50 mM Bicine-NaOH, pH 9.0, 37 OC, was mixed into the solution. The increase in fluorescence (Aexcitation = 360 nm, Xemission = 500 nm) was continuously
7328
Biochemistry, Vol. 28, No. 18, 1989
monitored at 37 OC. Activities were measured in fluorescence units per time, with a fluorescence unit defined as the maximum increase in dansylcadaverine fluorescence due to factor XIIIa activity under the assay conditions [ 1 fluorescence unit (FU) = fluorescence of a completely reacted sample fluorescence of a blank with 200 pL of buffer substituted for 200 pL of the factor XI11 containing solution]. Slopes of linear plots of FU (FU < 0.1) vs time were divided by the amount (milligrams) of factor XI11 in the assay cuvette to yield the specific activity for factor XIIIa (7FU min-' mg-I for pure factor XIIIa). Human a-thrombin was a generous gift of Dr. John W. Fenton I1 of the New York State Department of Health, Albany, NY. The thrombin was >98% a-thrombin and had an active-site content of 96%. Kinetics of A P release were measured by HPLC as previously described (Janus et a]., 1983). The reaction buffer contained 0.1 M Tris-acetate, 0.15 M NaCl, 1 mM EDTA, and 0.1% PEG all at pH 7.5 (TANEP). In some determinations (as specified), the reaction buffer also contained 14 mM CaC1,. Kinetics for the generation of platelet factor XIIIa activity were measured by using an adaptation of the dansylcadaverine incorporation assay (Curtis & Lorand, 1976; Curtis et al., 1974). For each time point, reaction mixtures were prepared consisting of 0.10 pM platelet factor XIII, 14 mM CaCl,, 20 nM a-thrombin, and 100-1 15 pL of TANEP to bring the volume to 125 pL. (The CaC1, was added immediately prior to the addition of a-thrombin to minimize a-thrombin-independent calcium activation of factor XIII.) At the indicated activation times, 10 pL of 30 pM PPACK (D-phenylalanylL-prolyl-L-arginine chloromethyl ketone, Calbiochem) was added to quench the thrombin activity. In control experiments, the resulting concentration (2.2 pM) of PPACK was shown to inhibit the a-thrombin activity within a few seconds without affecting the factor XIIIa activity, which was immediately measured by transferring a portion of the quenched reaction mixture into a waiting assay solution equilibrated in a fluorescence cuvette at 37 OC. Fluorescence activity was measured as described above. The fraction of maximal activity was determined by dividing the factor XIIIa activity observed by the maximal activity resulting from rapid activation with 220 nM a-thrombin. Kinetics of active-site thiol exposure of platelet factor XI11 were measured as previously described (Curtis et al., 1974). An incubate consisting of platelet factor XIII, a-thrombin, and 14 mM CaC1, in TANEP was prepared and treated as described above for measurements of the kinetics of generation of factor XIIIa activity. At specified time intervals, aliquots were pipetted into vials containing, as specified in the legends to Figures 3 and 4, amounts of PPACK to inactivate athrombin and amounts of [ l-'4C]iodoacetamide (New England Nuclear, 24.1 mCi/mmol) to alkylate exposed active-site thiol of factor XIIIa. Control experiments showed that the presence of PPACK at these specified concentrations did not affect the ability of iodoacetamide to alkylate factor XIIIa. After reaction with [1-14C]iodoacetamideat 37 OC for 12 or 30 min, aliquots were applied to Whatman 3MM filter paper (1 cm2), washed and dried as previously described (Curtis et al., 1973), and counted in 2.5 mL of scintillation cocktail (Bio-Safe 11, Research Products International Corp.) in a Beckman LS 3801 liquid scintillation system. Assay of a-Thrombin-Mediated Cleavage of Factor XIIIa. Cleavage of factor XIIIa by a-thrombin was assessed in a manner similar to that described previously (Takahashi et al.,
Hornyak et al. 1986). Platelet factor XI11 (1.0 pM) in TANEP was activated by 20 nM a-thrombin at 37 OC in the absence of and in the presence of 1.5 or 14 mM CaC1, for 0 min, 4 min (1.5 mM CaCl, only), 30 min, and 2 h. Aliquots (20 pL) were quenched at the indicated times by addition of 2 pL of 30 pM PPACK, after which 5 pL of a solution comprised of 0.15 M Tris-HC1, pH 6.8,40% glycerol, 5% SDS, 0.0075% BPB, and 125 mM DTT was added. The quenched samples were heated at 95 OC for 10 min and immediately transferred to a dry iceethanol bath where they remained until gel electrophoresis. Samples (9.5 pL) were applied to an 11.3% polyacrylamide gel prepared with an acrylamide:bis(acrylamide) ratio of 30:0.8 (Laemmli, 1970). The lanes were examined for the presence of the intact factor XI11 a subunit (MI SOOOO), the a'subunit ( M I 76 000) that results upon a-thrombin-catalyzed release of AP, and the a'subunit fragment ( M , 56 000) that results from a second a-thrombin-catalyzed cleavage of factor XIIIa (Schwartz et a]., 1973). RESULTSAND DISCUSSION Upon treatment with a-thrombin, platelet factor XI11 yielded two activation peptides, AP and AP', which were quantified by HPLC. As observed previously for plasma factor XIII, the yield of AP' was
